Time Series

Time Series Analysis is about the statistical analysis of data that is observed through time. The key aims of time series are modelling and forecasting. Most of my research has focused on locally stationary time series which is of importance in applications such as energy, epidemiology, and economics where it is increasingly being realized that underlying environments are not constant (the world changes!). I am also interested in sample rate problems.

More...

Applications

Much of my work is mathematical but my main aim is to design useful techniques for the real world. Hence, much of my research is translated to free easy-to-use software.

Current collaborators include EDF Paris, Shell, Vattenfall, Garrad Hassan, Office for National Statistics, and the Central Bank of Cyprus.

Past projects have been carried out with Barclays Bank, Fugro GEOS, the National Audit Office, National Air Traffic Services, Ministry of Defence (DERA, QinetiQ), Nuclear Electric, SAIC UK, Shell, Unilever and Windforce.

My work has been used by several companies and government agencies such as AC Nielsen, the Bank of America,the IMF, the Central Bank of Cyprus, Mathworks and the Reserve Bank of New Zealand.

Multiscale Methods

Many systems can profitably be examined at multiple scales simultaneously. Wavelets and multiscale methods can be used in many statistical areas such as regression, density estimation, survival data analysis, variance stabilization and time series. Multiscale methods are fast and highly efficient.

More ...

Network Time Series

Many important problems involve data on networks taken through time. For example, the evolution of a disease through a population. My key aim here is to forecast the progression of a disease using the strong information given by the underlying network structure that is often not used in classical approaches. Network forecasts can often perform much better than methods that do not use the structure.

More ...

Publication Info

List of publications

Google Scholar Profile

Applications

Most of my current interests are in official and economic statistics, computational politics, web scraping and epidemiology.